Potential issue with eparse mode

[luca-simonetti]

I don't fully understand if this is how modes other than eparse and structured (namely digest and table-digest) could ever work.
Was it supposed to be an elif to the eparse condition?

eparse/eparse/contrib/unstructured/xlsx.py

Lines 105 to 128 in 949d298

	if eparse_mode == "eparse":
	text = str(table.iloc[:eparse_max_rows, :eparse_max_cols])
	if "digest" in eparse_mode:
	digest = get_table_digest(
	df_serialize_table(table),
	table_name=table_name,
	)
	if eparse_mode == "table-digest":
	text = (
	f"{table_name} is a spreadsheet table. This is "
	f"the head of the table:\n{table.head(eparse_max_rows)}\n"
	f"Summary: {digest}."
	)
	else:
	text = digest
	elif eparse_mode == "unstructured":
	text = lxml.html.document_fromstring(html_text).text_content()
	table = Table(text=text, metadata=metadata)
	elements.append(table)
	return elements

[ChrisPappalardo]

Hi, thanks for opening this issue and your interest in eparse. The contrib function you’re referencing needs to be reworked as it was written for an old version of unstructured. You can either write your own control loop to identify and process xlsx files or change lines 105 and 113 to use in instead of ==. I’ll update this contrib module in a future release and fix these things in the update to latest unstructured.

[luca-simonetti]

thanks for your reply. I tried with a change on your code putting the ifs in the right order (or at least what I think it was the right one) and I think I actually need the vanilla eparse mode so I think I'll wait for the next release when there'll be one.